EVALUATION OF IMPROVED SHEAR KEY DESIGNS FOR MULTI-BEAM BOX GIRDER BRIDGES

Laboratory tests were conducted to investigate the problem of shear key failure in multi-beam prestressed box girder bridges and to propose a new shear key design. Failure of shear keys will typically not only compromise the load-sharing mechanism between adjacent girders, but also may lead to the failure of the deck waterproofing system, with attendant corrosion problems. The tests consisted of monitoring relative displacements occurring across intergirder joints, produced by a simulated concentrated wheel load. In order to economically conduct a reasonable number of tests, a 2-D "slice" of a multi-beam bridge cross section served as the basic test specimen. Finite element analyses of a 3-D bridge model showed that tensile stresses from transverse negative moments in the top flange of the bridge, generated as a result of the continuity provided by the current shear key design, may be responsible for many shear key failures. A proposed shear key solution is presented in a new location at the neutral axis of the box girder. Finite element analyses for the new shear key location indicate an elimination of the stresses suspected of causing the failure problem. Two series of lab tests were conducted to test a 2-D representation of the box girder bridge both under static and fatigue load. A 2-D representative "slice" was examined by finite element analysis to define the boundary conditions required to approximate the 3-D behavior. A total of three lab tests were conducted for the current shear key design as well as the proposed new shear key under static loading until failure of the shear key. All tests were repeated for three different grouting materials; non-shrink grout, mag-phosphate grout and epoxy grout, except that some of the test results could be anticipated from previously completed tests. A similar lab test program was repeated for fatigue life testing. Two reaction frames were utilized, each equipped with a 50 kip (223 kN) capacity actuator and a signal generator and controller system. A computer-based data acquisition system monitored a set of DCDTs and foil strain gages during each test. Grouting materials from individual test specimens were tested for tensile strength. Load, deflection, flexural strain and fatigue life were monitored and recorded. A final series of tests was run, complete with a waterproofing membrane and asphalt concrete overlay in place, to evaluate the effect of the modified shear key design on watertightness of the longitudinal joints. The proposed shear key, in the new neutral axis location, greatly improved the load-carrying capacity of the tested specimens when compared to the current shear key design. Utilizing a current ODOT "Type III" waterproofing membrane, at laboratory temperatures in a non-chloride environment, watertightness was maintained, even in the presence of failed shear keys of the "current" design.